Calculating Limits of Base Station Emission Power in GSM...Calculating Limits of Base Station Emission Power in GSM R. vraka, D. Miti c, A. Lebl, . Markov´ where N no is the number

Rade Švraka, Dragan Mitic, Aleksandar Lebl, Žarko Markov

Calculating Limits of Base Station Emission Power in GSM

DOIUDK

10.7305/automatika.2017.02.994621.396.7-531.9:621.391(083.74)GSM; 620.9(083.744)GREEN

Original scientific paper

This paper compares the influence of various factors on the output power of each GSM channel and connection,and on the output power of whole base transceiver station in GSM network. The considered factors are the influenceof sending signal only when speech exists (DTX), implementation of power control in signal sending, influence ofusers’ distribution in the range of coverage of base station, the power of one intra-cell connection and the powerof one half-rate connection. It is analyzed how emission power of base station can be reduced by implementingvarious methods in signal sending.

Key words: base transceiver station, GSM channel, influence of various factors on output power

Izracunavanje granica emisije snage bazne stanice u GSM-u. U ovom radu uspore�uje se utjecaj razlicitihcimbenika na izlaznu snagu svakog GSM kanala i veze te utjecaj na izlaznu snagu cijele bazne stanice u GSMmreži. Promatrani cimbenici su utjecaj predajnog signala samo kada govorni signal postoji (DTX), implementacijakontrole snage pri slanju signala, utjecaj raspodjele korisnika ovisno i podrucju koje pokriva bazna stanica, snagajedne unutar-celijske veze i snaga veze koja koristi pola protoka. Analizirano je kako se emitirana snaga baznestanice može smanjiti koristeci razlicite metode za slanje signala.

Kljucne rijeci: bazna stanica, GSM kanal, utjecaj razlicitih cimbenika na izlaznu snagu

1 INTRODUCTION

The usage of radio communication is greatly increas-ing. That’s why it is concluded that energy saving is im-portant at least because of two main reasons. The first oneis price of energy, and the second one is environmental pro-tection, [1], [2]. Energy saving is important at all elements,included in wireless communication. At remote units, it isimportant, because they are powered by batteries, whichhave limited capacity [3]. The scenario of energy sav-ing is called GREEN (Globally Resource-optimized andEnergy-Efficient Networks) Radio. In order to implementthis program, it is very important to know the dependenceof BTS (Base Transceiver Station) output power on vari-ous factors. Energy consumption, i.e. instantaneous elec-trical power consumption of BTS in the network of mobiletelephones (GSM) depends on various factors, as trafficvalue, signal attenuation in GSM cell, users’ density dis-tribution in the cell, implementation of half rate channels,part (percent) of intra-cell connections. When the trans-mitting power decreases, self-interference also decreases,and so the necessary quality of service is achieved (Qualityof Service, QoS) in the conditions of optimal BTS outputpower. Considering all these reasons, it is necessary toknow the dependence of BTS output power on all factors

considered in the analyzed cell. In this paper we shall indi-cate the limits for the values of output power in one cell. Inthe section 2 the model, abbreviations and some assump-tions used in the paper are presented. In the section 3 itis described how power comparison is made. In sections4-7 the influence of each characteristic on output power ofchannel and BTS is analyzed. In the last sections compar-ison of output power of different models is performed.

2 MODEL, DESIGNATIONS AND ASSUMPTIONS

Let us consider one BTS and one cell in GSM network,[1], [2]. The cell is circular and its radius is R. The networkfunctions using FDMA (Frequency Division Multiple Ac-cess) and TDMA (Time Division Multiple Access) is usedon each carrier (frequency). The number of carriers, thatare used in one BTS, is n and the number of time slotson each frequency is Ns(=8). Number of mobile stations(MS) is Nms and it is much greater than the number oftraffic channels N . The mean number of active channelsis designated as Nac. In this paper we consider the caseswhen the output power of one connection from BTS is notchanged (cell without power control), or it is changed ac-cording to the needs of the concrete connection (cell withpower control). Half rate (HR) connections can be used

Online ISSN 1848-3380, Print ISSN 0005-1144ATKAFF 57(3), 774–781(2016)

774 AUTOMATIKA 57(2016) 3, 774–781

Calculating Limits of Base Station Emission Power in GSM R. Švraka, D. Mitic, A. Lebl, Ž. Markov

in the cell. The model with power control is considered.It is supposed that even in this model on some channelspower control is not implemented (e.g. channels on thefirst carrier). The part of connections between the users be-longing to the same cell (intra-cell (internal) connections,Fig. 1) can be very great, especially when one BTS coversone place (town) or great company [4], [5]. The distribu-tion of users in the cell can be very different: starting fromuniform distribution to variable distribution, which can in-crease from the centre of the cell towards the rim, or viceversa. The offered traffic to the BTS is A. The call loss,caused by the lack of free traffic channels, is designated byB. The served traffic, i.e. mean number of realized con-nections, is designated by Y , Y = A · (1 − B) and it is,generally, different from the mean number of busy chan-nels, Nac, because two channels are necessary for intra-cell connection, or two HR connections can be realized onone channel.

BTSR

MS4MS3

MS1

internalMS2external

Fig. 1. Schematic presentation of internal and externalconnections

Instantaneous, mean and maximum power of one chan-nel are designated by w1, w1m and w1max. Mean power,which is necessary for one connection, is designated byw1cm. Maximum power corresponds to one of classes [1],and it is supposed that this power is constant. Output powerof BTS is wB . The BTS output power is equal to oneeighth of the sum of output power of each active channel,ω1 = (1/8) · w1. The traffic channel, which uses the fullrate, is designated by FR and the channel, which uses halfrate, is designated by HR. Internal connections are estab-lished between the users, who are situated in the area of thecoverage of the same BTS. On the schematic presentationin the Fig. 1, internal connection is realized between usersMS3 and MS4. On the other side, external connection is

realized between the users from the areas of two differentusers (MS1 and MS2 in Fig. 1). Throughout the paper, in-ternal connections are designated by an index i and exter-nal connections (outgoing and incoming connections) aredesignated by index e. It is supposed that attenuation is thesame in the whole cell and that the cell dimensions dependon the attenuation value (when attenuation is increased, thecell radius is decreased and vice versa).

3 HOW POWERS ARE COMPARED

In this paper we shall estimate the influence of differentfactors on BTS output power. These factors can be dividedin two groups: the group, which has influence on the out-put power of each GSM channel and the group, which hasinfluence on total BTS output power.

In the first group are: transmission of only parts withspeech signal (i.e. pauses in speech are not transmitted(DTX), propagation attenuation, distance MS–BTS (the in-fluence is realized by the distribution of users in the cell)and in the second group: bit flow used for the transmis-sion of one telephone connection (HR, FR) and the kind ofconnection (internal, external).

Comparative value of power, w1max, which is the max-imum value of one channel output power necessary for thetransmission of speech using full rate (FR), is defined forthe comparison of output power in one channel.

4 POWER CONTROL AND DTX

As presented in Fig. 2, which is taken from [6], allBTSs can be divided in two groups: the ones which haveno output power control (Fig. 2.a)) and the ones, whichhave power control (Fig. 2.c)). BTSs, which have outputpower control, have no such control for the channels onthe first carrier f1 (Fig. 2.c)). Besides this, traffic channelson the first carrier always have maximum power, no matterthey are busy or not (channels 5 and 6, Fig. 2.).

If there is no power control, total BTS output power is:

wB =1

8· (8 +Nac) · w1max (1)

where Nac is here the number of all active channels in thecarriers f2, ..., fn.

One of the methods to decrease BTS output power isto use discontinuous transmission, DTX, Fig. 2.b). In thiscase speech signal is not transmitted during the pauses inthe speech. It can be supposed that BTS output power is,then, 60% of the power when DTX is not used, [7]:

wBDTX =1

8·(Nno + 0.6 · 1

8·Nac

)· w1max (2)

AUTOMATIKA 57(2016) 3, 774–781 775


where Nno is the number of all active channels where nopower control is implemented.

If power control exists, then mean power of one chan-nel and mean number of necessary channels for one con-nection must be calculated to determine BTS output power.Mean value of output power of one channel when powercontrol is implemented depends on propagation attenua-tion and on mean value of distance between MS and BTS.

w1max

012 76543 012 76543 012 76543

b)

f1 f2 fn

time slotscarrier

w1max

012 76543 012 76543 012 76543

c)

f1 f2 fn

time slotscarrier

w1max

012 76543 012 76543 012 76543

a)

f1 f2 fn

time slotscarrier

DTX

power control

control (non traffic) channels channel with dummy burstsbusy traffic channel

Fig. 2. Output power of BTSs without power control (Fig.2a)), with the implementation of DTX (Fig. 2b)) and withpower control (Fig. 2c))

5 PROPAGATION ATTENUATION

Propagation attenuation affects the output power of onechannel with power control according to the law:

w = a · dγ + b (3)

where a is coefficient of proportionality, γ = 2 − 5, [1],[2], and b is output power for the connections of users whoare in the immediate vicinity of BTS. This power b can beneglected in the following considerations, so w = a · dγ .As the greatest power of channel is constant, the radius ofcell can be determined from the equation wmax = a ·Rγ .

It is proven in [8] that mean output power of one chan-nel, w1m, if power control exists, and signal is transmittedin the environment where coefficient of propagation atten-uation is γ and users’ distribution in the cell is uniform,equals:

w1m =2 · w1max

γ + 2(4)

It can be concluded from (4) that the cell and the meanpower of one channel are smaller for the greater values ofpropagation attenuation. Although (4) is valid for the cellwith uniform distribution of users, the conclusion about theinfluence of coefficient γ can be spread for all distributionsof MSs density in the cell.

6 DENSITY

It is obvious from equation (3) that output power of onechannel in the cell with power control depends on distanceMS-BTS. Traffic process is random, so the distance MS–BTS, d, can be considered as independent random variable.The mean value of this variable depends on users’ positionin the cell, i.e. on distribution of users’ area density in thecell. It is clear that mean distance MS–BTS will be smallerif greater number of users is concentrated near BTS, thanin the case of uniform users’ distribution in the cell. It isalso clear that smaller mean distance MS–BTS will causesmaller mean value of output power of one channel andsmaller total power of BTS.

7 INTRA-CELL

Besides the considered factors, which have influenceon the mean value of output power of each channel, twomore factors have influence on the mean BTS outputpower: part of intra-cell connections, [4] and [9], and theuse of one channel for two connections (half rate), [10] -[12].

Intra-cell (or internal) connections in the cell alwaysexist, but in different percent. This percent is negligible inthe town, but in one place, which is covered by only oneBTS can be significant [4], [5]. The part of intra-cell con-nections can be very great in the cells covering rural area ora great company. The part of intra-cell connections will bedesignated by p. Intra-cell connection occupies two chan-nels, while external connection occupies only one channel[13].

The mean output power for one intra-cell connection is

w1cmi = 2 · w1cme (5)

The mean BTS output power for external connectionscan be calculated as the product of mean output power ofone channel and mean number of active channels for exter-nal connections. The mean number of active channels forexternal connections is equal to the value of served exter-nal traffic

776 AUTOMATIKA 57(2016) 3, 774–781


wBem = ω1m ·Nace == ω1m · Ye = ω1m · (1−Be) ·Ae (6)

The losses are often negligible, Be ≈ 0, so it iswBem = ω1mAe.

The mean value of BTS output power for intra-cell con-nections can be calculated as the product of mean outputpower of one channel and the mean number of active chan-nels. The mean number of active channels is equal to thedouble value of served internal traffic

wBim = ω1m ·Naci == 2 · ω1m · Yi = 2 · ω1m · (1−Bi) ·Ai (7)

or, often wBim = 2ω1mAi.It can be concluded that mean power of one chan-

nel, i.e. the relation of mean output power wBm =wBem + wBim and total offered traffic A = Ae +Aiequals (Ae + 2 ·Ai)/(Ae +Ai) and that it is as greateras the component of intra-cell traffic is greater.

8 HALF RATEIn order to reduce losses caused by the lack of channels

on Air-interface, dynamic channel allocation with half rate(HR) is often used, [1], [10], [11], [12]. Usage of HR con-nections by the implementation of repacking channel as-signment is considered in this paper. (Repacking schemeof HR channel assignment means that one time slot is usedfor two HR connections, [11], [12]). It is clear that meanoutput power of one channel w1m is divided on two con-nections. That’s why the mean output power of one HRconnection is equal to half the power of one FR connec-tion:

w1cmHR = 0.5 · w1cmFR (8)The mean BTS output power for certain number of con-

nections is as smaller as the part of HR connections isgreater. The probability that the connection is realized us-ing half rate is designated by q.

For GSM traffic realization, besides reduction of callloss probability and reduction of BTS emission power, itis very important to improve voice quality. Several ap-proaches are applicable to achieve this goal. Enhancedfull-rate codec (EFR) (also designated as GSM 06.60) isone method to obtain connection quality not much worsethan the quality obtained by G.711 codec, but with muchless throughput. The other method is implementation ofadaptive multirate voice codec (AMR), which can be real-ized as narrow-band (AMR-NB) or wide-band (AMR-WB)codec. The first one (AMR-NB) supports the bandwidth oftraditional telephony, while the other one (AMR-WB) sup-ports two-fold bandwidth (till 7 kHz). Some of AMR-NBcodecs can be implemented in HR channels, thus reducingBTS output power.

9 COMPARISON OF MODELS

The following statements can be expressed for the BTSoutput power in GSM network:

9.a. The greatest mean power, independently of powercontrol implementation, is the power in time slots, belong-ing to the first carrier. The output power of time slots infirst carrier is constant and, according to (1), its value isalways:

wB = w1max (9)

This power does not depend on the number of connec-tions, which are realized using time slots of first carrier. Ifmore than one carrier exists, output power depends also onthe traffic.

9.b. The output power of other carriers (f2, ..., fn) de-pends on the number of active channels. This output powercan be treated as the value of total output power and as thevalue of power of one connection.

9.b.1. The greatest value of output power is in themodel of BTS without power control, and it is equal tothe product of the number of active channels and the valueof greatest channel power, as in (1)

wB = Nac · w1max (10)

9.b.2. The output power is little less than in case 9.b.1.,if DTX is implemented, the power is 2.2dB smaller thanwithout DTX.

9.b.3. The output power, which is used for one connec-tion, is obtained as the quotient of mean BTS power, wBm,and the traffic (it is supposed that the loss is small and thatthe served traffic is equal to the offered traffic), A. Themean power is obtained if the connections, which use onechannel, two channels and half of the channel, are consid-ered. This value depends on the part of intra-cell traffic (p)and on the part of HR connections (q). One channel is usedby:

- external connections with full rate, their probability is(1-p)·(1-q) and

- internal connections with HR, their probability is p·q.Intra-cell connections with full rate use two channels,

and their probability is p·(1-q).Half channel is used by external connections with half

rate, the probability of these connections is q·(1-p).The mean BTS power without power control is

wBm =

(1− p) · (1− q) + p · q++2 · p · (1− q)++0.5 · q · (1− p)

·A · w1max

(11)

AUTOMATIKA 57(2016) 3, 774–781 777


and mean power of one connection is

w1cm =

(1− p) · (1− q) + p · q++2 · p · (1− q)++0.5 · q · (1− p)

· w1max (12)

It is obvious that mean BTS output power increaseswhen p increases and decreases when q increases.

9.b.4. Further power reduction is possible if imple-menting power control. In that case the mean BTS outputpower is equal to the product of mean number of activechannels and mean value of power of one channel.

The greatest mean power is now obtained when signalis transmitted with the smallest loss, γ=2. For uniformusers’ distribution in the cell, this mean power has half thevalue of power without power control (3dB less than thepower in case 9.b.1.), [8]. In real case, when γ=3-4, thepower is decreased between 4dB and 4.8dB.

It is obvious that the power for one connection in thiscase also increases when increasing the part of intra-cellconnections, and that it decreases when increasing thenumber of HR connections. The equation (12) for BTSwith power control becomes:

w1cm =

(1− p) · (1− q) + p · q++2 · p · (1− q)++0.5 · q · (1− p)

· w1m (13)

9.b.5. The smallest power is obtained when the co-efficient of loss is great, (γ=5, reduction of 5.4dB) andwhen the great number of users are concentrated in thenear vicinity of BTS. It is presented by the results of mea-surements in [7], Figure 9, that implemented power controlin some cases can decrease output power in the range 10–12dB. In some cases, where we suppose that users’ concen-tration around BTS is great, BTS output power reductioncan reach 26dB.

Figure 3 presents graphically the reduction of meanpower of one channel in the GSM cell, depending on: DTXimplementation, power control usage, coefficient of envi-ronment attenuation and distribution of users in the cell.Points 1–6 in Fig. 3 represent the following cases:

- point 1 represents the power of one channel whenthere is no power control and no DTX implementation;

- point 2 represents the case of channel power withoutpower control, but with DTX implementation;

- point 3 represents mean output power of one chan-nel in the cell with uniform distribution of users and ideal(minimum) value of γ=2;

- point 4 corresponds to the mean output power of onechannel in the cell with uniform users’ distribution andmedium value γ=3-4;

- point 5 represents the mean output power of one chan-nel in the cell with uniform users’ distribution and withgreat value γ=5;

- point 6 corresponds to the mean output power of onechannel in the cell where the users’ density decreases lin-early from BTS to the rim of the cell [6] (it is six timessmaller on the rim than in the centre), and where it is, γ=5.

It is necessary to remark that mean output power of onechannel can be decreased even more, if the users’ densitynear BTS is greater than it is presented for point 6.

Figure 4 represent the range of mean output power perone connection. The greatest value of output power fromFig. 3 can be doubled for one connection in the case ofintra-cell connection, point 1 in Fig. 4. The mean powerof one connection can be also halved if the connection isexternal and if it is realized using HR, point 5 in Fig. 4.

10 PROCESS OF SIMULATION

The values presented in Figures 3 and 4 and the otherconclusions from this paper are proved by simulation pro-cess, which is explained in more detail in [13], [14]. Themain goal of the simulation was to verify values of meanpower of one connection for various users’ density distri-butions and for various coefficients of attenuation. Theprocess of simulation is performed in such a way that, foreach simulation run, users’ distribution and coefficient ofattenuation are chosen at first. In order to simulate realsituation, 8, 16 or even more channels exist in the sim-ulation and at least one thousand of calls are generatedper each channel. Real traffic process is also simulated,which means that we simulated calls generation and ter-mination. The mean output power of one channel is thandetermined as the mean power for all connections realizedin each channel, considering also call duration.

The simulation is based on Monte Carlo, or roulettemodel. The generated random numbers are uniformly dis-tributed in the range (0, A+N), where A is traffic load, of-fered to the simulated group of channels and N is the num-ber of channels. In the case that generated random numberfalls in the range (0, A), the new call is generated (if lessthan N channels are busy in that moment). The range ofnumbers (A, A+N) is divided in N parts (subranges), eachsubrange corresponding to one traffic channel. In the casethat generated random number falls in the subrange (A+i,A+i+1), the call is terminated if HR channel i+1 is busyin that moment. In such way traffic part of simulation isperformed.

Determination of user’s distance from BTS is, gener-ally, more complicate. It is because we want to simulatesome areal distribution of users, and we need to calculatelinear distance between MS and BTS. Thus the problem

778 AUTOMATIKA 57(2016) 3, 774–781


w1max0.6*w1maxw1m(γs,uu)w1m(γav,uu)w1m(γg,uu)w1m(γg,uc)

123456

3dB2.4dB2.6dB

DTXWPCPC

PC power control, WPC without power control

1. w1max - no power control, no DTX implementation

2. 0.6*w1max - no power control, DTX implementation

3. w1m(γs,uu) - γ of small value (2), users uniformly distributed,

4. w1m(γav,uu) - γ of average value, users uniformly distributed,

5. w1m(γg,uu) - γ of great value (5), users uniformly distributed,

6. w1m(γg,uc) - γ of great value, users concentrated near BTS

Fig. 3. Symbolic presentation of the relation of mean one channel power in the cell of GSM network depending on thepower control, coefficient of attenuation and distribution of users in the cell

w1cmi=2·w1maxw1cme=w1max0.5·w1maxw1cmFR(γg,uc)w1cmHR(γg,uc)12345

3dBpower decrease per channel (figure 3)3dB

DTXWPCPC

1. Power of intracell connection without power control

2. Power of external connection without power control

3. Power of external HR connection without power control

4. Power of external connection in the cell with users

concentrated around BTS, great coefficient of attenuation and

with power control

5. Power of external HR connection in the cell with users

concentrated around BTS, great coefficient of attenuation and

with power control

Fig. 4. Symbolic presentation of the relation of mean power of in the cell of GSM network depending on the type ofconnection, power control, coefficient of attenuation and distribution of users in the cell

of finding distance between MS and BTS leads to find-ing inverse function of the function, which models users’density. If the users are uniformly distributed, the calcu-lation of distance between MS and BTS is based on cal-culating square root of generated, uniformly distributed,random number. But, in general case, it is not ease to de-termine inverse function in closed form, and we use pro-grams MATHEMATICA or MATLAB to calculate it. Ineven more complicate case, there is no closed form expres-sion for inverse function. Then we determine approximateexpression for inverse function, again using MATHEMAT-ICA or MATLAB. In such a way we can determine power

of the connection, when users are concentrated aroundBTS, Fig. 4.

Calculation of BTS power, depending on signal atten-uation, is performed by using distance between MS andBTS, which is calculated according to one of methods ex-plained in the previous paragraph. The calculated value ofthis distance is introduced to the formula, corresponding tothe simulated attenuation law.

The program, which we developed, can, also, simulateinfluence of intra-cell connections. The part of intra-cellconnections in total number of connections can be change-able. But, in the simulation relevant for this paper we used

AUTOMATIKA 57(2016) 3, 774–781 779


0% of intra-cell connections (complete traffic is external)and 100% of intra-cell connections (there is no externaltraffic, i.e. complete traffic is intra-cell).

11 CONCLUSION

In this paper we indicated some factors, which have in-fluence on the increase or decrease of the power of oneBTS. In this paper we have tried to determine the possi-ble saving of transmitted power in one typical BTS. Theoutput power of traffic channels in one GSM BTS can bedecreased using DTX technique, but maximum reductionis achieved by implementing power control, except for thechannels on the first carrier. It can be seen from the exam-ple that the reduction can be till 10dB. In some cases, [7],it can be even greater. The degree of saving depends on thenumber of carriers: if there are more carriers, power sav-ing is greater. If users are concentrated near BTS, powerreduction is significant if power control is implemented.

The power, needed for one connection, has even greaterrange if the part of intra-cell connections is great. If thepart of intra-cell connections is great, BTS must be sup-plied by greater power then if the number of external con-nections is dominant. Mean power per one connection canbe reduced using the decreased flow for one connection(HR).

REFERENCES

[1] H. Gunar, GSM Networks: Protocols, Terminology and Im-plementation, Artech House, 1999.

[2] J. Ebersprächer, H.-J. CityplaceVögel, country-regionCh.Bettstetter,: GSM, Switching, Services and Protocols, JohnWiley & Sons, 1999.

[3] R. Klempous, “Collective Behaviour in Wireless SensorNetworks,” Acta Polytechnica Hungarica, vol. 11, no. 4,pp. 101-118, 2014.

[4] P. Jovanovic, T. Šuh, A. Lebl, D. Mitic, Ž. Markov, “Influ-ence of Intra-cell Connections on the Traffic Calculationsof Radio Resources in Mobile Network,” Frequenz, vol. 67,issue 9-10, pp. 315-320, 2013.

[5] A. Anand, V. Pejovic, E. M. Belding, D. L. Johnson, “Vil-lageCell: Cost Effective Cellular Connectivity in Rural Ar-eas”, Proceedings of the Fifth International Conference onInformation and Communication Technologies and Devel-opment, ICTD’12, placeCityAtlanta, 2012.

[6] M. Mileusnic, M. Popovic, A. Lebl, D. Mitic, Ž. Markov,“Power Control of Base Station in GSM: Influence ofUsers’ Density in the Cell,” Automatika, vol. 56, no. 1, pp.84-90, March 2015.

[7] D. Colombi, B. Thors, T. Persson, N. Wiren, L.E. Larsson,M. Jonsson, C. Tomevik, “Downlink Power Distribution for2G and 3G Mobile Communication Networks,” RadiationProtection Dosimetry, vol. 157, no. 4, pp. 477-487, 2013.

[8] P. Jovanovic, M. Mileusnic, A. Lebl, D. Mitic, Ž. Markov,“Calculation of the Mean Output Power of Base TransceiverStation in GSM,” Automatika, vol. 55, no. 2, pp. 182-187,2014.

[9] T. Šuh, P. Jovanovic, A. Lebl, D. Mitic, Ž. Markov, “Com-parison of the Influence of Intra-cell Traffic and Finite Num-ber of Mobile Phones on the Determination of Number ofChannels in the BTS of GSM Network,” Frequenz, vol. 68,issue 3-4, pp. 171-176, 2014.

[10] M. Ivanovich, M. Zukerman, P. Fitzpatrick, M. Gitlits, “Per-formance Between Circuit Allocation Schemes for Half-and Full-Rate Connections in GSM,” IEEE Transactions onVehicular Technology, vol. 47, no. 3, pp. 790-797, 1998.

[11] P. Lin, Y. B. Lin, “Channel Assignment for GSM Half-Rateand Full-Rate Traffic,” Computer Communications, vol. 23,no. 5-6, pp. 476-482, 2000.

[12] E. M. Winands, B. Wieland, J. Sanders, “Dynamic Half-rateConnections in GSM,” AEÜ, vol. 60, no. 7, pp. 504-512,2006.

[13] M. Mileusnic, P. Jovanovic, M. Popovic, A. Lebl, D. Mitic,Ž. Markov, “Influence of Intra-cell Traffic on the OutputPower of Base Station in GSM,” Radioengineering, vol. 23,no. 2, pp. 601-608, 2014.

[14] M. Mileusnic, T. Šuh, A. Lebl, D. Mitic, Ž. Markov, “Useof Computer Simulation in Estimation of GSM Base StationOutput Power,” Acta Polytechnica Hungarica, vol. 11, no.6, pp. 129-142, 2014.

Rade Švraka was born in Zagreb, Croatia in1988. From 1992 he lives in Belgrade. He en-tered University of Belgrade, School of Electri-cal Engineering in 2007. He graduated from De-partment of Telecommunications and Informa-tion Technologies in 2011. He got MSc in Sys-tem Engineering and Radio Communications atBelgrade University, School of Electrical Engi-

neering, Department of Telecommunications and Information Technol-ogy. Rade joined Iritel (Institite for Telecommunications and Electron-ics) in 2012 and since then he works as an electrical engineer in Depart-ment for Optical Transmission Systems. His main area of interest andresearch is designing and management of OTN (Optical Transport Net-work) and SDH (Synchronous Digital Hierarchy) networks. He speaksSerbian, Russian, English and German.

Dragan Mitic was born in Belgrade, Serbia, in1953. He received his B.Sc. and M.Sc. fromthe Faculty of Electrical Engineering in Belgrade,Republic of Serbia, in 1977. and 1984., respec-tively, and his Ph.D. from the Faculty of Tech-nical Science in Novi Sad, in 2002. Dr Miticis a scientific associate in IRITEL, Institute forElectronics and Telecommunications, Belgrade,Serbia. From 1977 until 1989 he was employedat the Land Forces Military Technical Institute in

Belgrade, and since 1989 in IRITEL. Author or co-author of more than70 international and national scientific and professional papers. He workson several research projects for equipment of specific applications.

780 AUTOMATIKA 57(2016) 3, 774–781


Aleksandar Lebl was born in Zemun, Serbia, in1957. He received his B.Sc. and M.Sc. fromthe Faculty of Electrical Engineering in Belgrade,Republic of Serbia, in 1981. and 1986., respec-tively, and his Ph.D. from the Faculty of Techni-cal Science in Novi Sad, in 2009. He is employedfrom 1981 in the Switching Department of Insti-tute for Electronics and Telecommunications IRI-TEL in Belgrade. During years he worked on theproject of Digital Switching System for Serbian

Telecommunication Industry.

Žarko Markov was born in Žitište, Serbia, in1946. He received his B.Sc., M.Sc. and Ph.D.from the Faculty of Electrical Engineering in Bel-grade, Republic of Serbia, in 1969., 1975. and1976., respectively. Dr Markov is a scientificcounsellor in IRITEL, Institute for Electronicsand Telecommunications, Belgrade, Serbia. Areaof work: Switching technics, Teletraffic theory,Network signalling. Author or co-author of morethan hundred papers and six books. At the Uni-versity of Belgrade, School of Electrical Engi-

neering, Dr. Markov was a professor at the course of Switching technicsand Network signalling.

AUTHORS’ ADDRESSESRade Švraka, M.Sc.Aleksandar Lebl, Ph.D.Dragan Mitic, Ph.D.prof. Žarko Markov, Ph.D.IRITEL AD Beograd, Batajnicki put 23, 11080 Belgrade,Serbia,email: [email protected], [email protected],[email protected], [email protected]

Received: 2014-09-18Accepted: 2014-09-18

AUTOMATIKA 57(2016) 3, 774–781 781

Calculating Limits of Base Station Emission Power in GSM...Calculating Limits of Base Station Emission Power in GSM R. vraka, D. Miti c, A. Lebl, . Markov´ where N no is the number

Documents